检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Salvatori, Tommaso Millidge, Beren Song, Yuhang Bogacz, Rafal Lukasiewicz, Thomas VERSES AI Research Lab Los AngelesCA90016 United States Medical Research Council Brain Network Dynamics Unit University of Oxford United Kingdom Institute of Logic and Computation Vienna University of Technology Austria Department of Computer Science University of Oxford United Kingdom

An autoassociative memory model is a function that, given a set of data points, takes as input an arbitrary vector and outputs the most similar data point from the memorized set. However, popular memory models fail to retrieve images even when the corruption is mild and easy to detect for a human evaluator. This is because similarities are evaluated in the raw pixel space, which does not contain any semantic information about the images. This problem can be easily solved by computing similarities in an embedding space instead of the pixel space. We show that an effective way of computing such embeddings is via a network pretrained with a contrastive loss. As the dimension of embedding spaces is often significantly smaller than the pixel space, we also have a faster computation of similarity scores. We test this method on complex datasets such as CIFAR10 and STL10. An additional drawback of current models is the need of storing the whole dataset in the pixel space, which is often extremely large. We relax this condition and propose a class of memory models that only stores low-dimensional semantic embeddings, and uses them to retrieve similar, but not identical, memories. We demonstrate a proof of concept of this method on a simple task on the MNIST dataset. Copyright © 2024, The Authors. All rights reserved.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

On TinyML WiFi Fingerprinting-Based Indoor Localization: Comparing RSSI vs. CSI Utilization

On TinyML WiFi Fingerprinting-Based Indoor Localization: Com...

引用

Consumer Communications and Networking Conference, CCNC IEEE

作者： Diego Mendez Marco Zennaro Moez Altayeb Pietro Manzoni Electronics Engineering Department School of Engineering Pontificia Universidad Javeriana Bogotá Colombia Science Technology and Innovation Unit The Abdus Salam International Centre for Theoretical Physics Trieste Italy Department of Computer Engineering School of Informatics Universitat Politècnica de València Valencia Spain

As context-aware location-based services (LBS) become increasingly important in many Internet of Things (IoT) verticals, such as logistics or industry 4.0, indoor localization is now an essential feature to be integrated in these solutions. For this purpose, fingerprinting-based solutions arise as a feasible solution, especially when integrating artificial intelligence on the edge, supported by computational and memory-restricted embedded devices, as it does not depend on a cloud-based deployment. In this work, we integrate this new paradigm, known as TinyML, and compare the implementation of a machine learning (ML) model when using only WiFi Received Signal Strength Indicator (RSSI) or WiFi Channel State Information (CSI) data. We tested two different scenarios, a single sample or time series, with different configurations of the trained neural network. Our results show that a CSI data ML model always outperforms an equivalent RSSI approach, with a massive difference in performance for the time-series case.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SKIN_ML: An Efficient Approach for Skin Cancer Detection Using Soft Computing Methods

SKIN_ML: An Efficient Approach for Skin Cancer Detection Usi...

引用

Mobile, Intelligent, and Ubiquitous Computing Conference (MIUCC), International

作者： Diaa Salama AbdElminaam Magda M. Madbouly Mohamed S Farag Ibrahim A Gomaa Magdy Abd-Elghany Zeid Laith Abualigah Faculty of Computer Science Misr International University Cairo Egypt Faculty of Computers and Artificial Intelligence Benha University Benha Egypt Faculty of Computer and Data Science Alexandria University Alexandria Egypt Computer Science Department Alobour high institute for computer and informatics Cairo Egypt Computer Science Department Al al-Bayt University Mafraq Jordan Department of Electrical and Computer Engineering Lebanese American University Byblos Lebanon Hourani Center for Applied Scientific Research Al-Ahliyya Amman University Amman Jordan MEU Research Unit Middle East University Amman Jordan School of Computer Sciences Universiti Sains Malaysia Pulau Pinang Malaysia School of Engineering and Technology Sunway University Malaysia Petaling Jaya Malaysia

In today's environment, cancer is a fatal disease. Skin cancer has become a fairly common malignancy due to the spread of several forms of cancer. Skin cancer is divided into two types: melanoma and non-melanoma. Melanoma is one of the most fatal tumors on the planet, and it can spread to other parts of the body if not diagnosed early enough. Our proposed system uses Five alternative methods to predict a skin lesion's borders, texture, and color: a neural network and four standard machine learning classifiers. To enhance their performance, the outputs of these systems are merged using majority voting. Experiments have demonstrated that combining the five strategies yields the maximum level of accuracy. Pre-processing, Segmentation, Feature Extraction, and Classification are four critical phases in skin cancer identification. Skin lesion images were collected for this research from the International Skin Imaging Collaboration (ISIC), which contains over 3297 photos. The accuracy of the NN classifier is 91.9%, compared to 87.2% for the KNN classifier, 81.5% for the Naive Bayes classifier, 72.5% for the SVM classifier, and 68.3% for the DT classifier.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ML_Alzheimer: Alzheimer Disease Prediction Using Machine Learning

ML_Alzheimer: Alzheimer Disease Prediction Using Machine Lea...

引用

Mobile, Intelligent, and Ubiquitous Computing Conference (MIUCC), International

Alzheimer's disease is a fatal brain disorder that impacts predominantly the elderly. The early identification of Alzheimer's illness requires the use of efficient automated methods. For the categorization of Alzheimer's disease, researchers have proposed several distinct methods. To create more effective learning strategies, however, a deeper comprehension of Alzheimer's research is necessary. The condition is diagnosed but not until a later stage. As a consequence, the disease's progression or its symptoms' onset can be slowed if detected earlier. This article uses two datasets and machine learning techniques to predict Alzheimer's disease based on psychological characteristics such as age, gender, number of visits, and Mini-Mental State Examination (MMSE). We ultimately focused on the masculine gender and characteristics. The following results (accuracy) were discovered: SVM: 84%, KNN: 79%, and Random forest: 86%. Neural network: 86%; Naive Bayes: 83%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ML_Recognition: A Robust Model for Handwritten Digit Recognition Using Machine Learning Classifiers

ML_Recognition: A Robust Model for Handwritten Digit Recogni...

引用

Mobile, Intelligent, and Ubiquitous Computing Conference (MIUCC), International

作者： Diaa Salama AbdElminaam Magda M. Madbouly Magdy Abd-Elghany Zeid Laith Abualigah Faculty of Computer Science Misr International University Cairo Egypt Faculty of Computers and Artificial Intelligence Benha University Benha Egypt Faculty of Computer and Data Science Alexandria University Alexandria Egypt Ibrahim Abd Elatif Gomaa Computer Science Department Alobour high institute for computer and informatics Cairo Egypt Computer Science Department Al al-Bayt University Mafraq Jordan Department of Electrical and Computer Engineering Lebanese American University Byblos Lebanon Hourani Center for Applied Scientific Research Al-Ahliyya Amman University Amman Jordan MEU Research Unit Middle East University Amman Jordan School of Computer Sciences Universiti Sains Malaysia Pulau Pinang Malaysia School of Engineering and Technology Sunway University Malaysia Petaling Jaya Malaysia

Handwritten digit recognition is a branch of machine learning in which a computer is taught to recognize hand-written numbers. Classification and regression are applied using deep learning and machine learning algorithms. In this paper, we discussed the efficiency of different algorithms: Random Forest, KNN, Naive Bayes, SVM, CNN, and Decision Tree Algorithm. Those algorithms were applied in the handwritten digit recognition process using the MINIST data set. By comparing each algorithm by training every one of them on the data set, then by testing, we got the best algorithm to get the optimal results with the highest accuracy. After comparing the results of each algorithm, it has shown that the neural network has the best results with 97.3% classification accuracy followed by K- the nearest neighbor with 97.2%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

InGaAs/InP Quantum Well Nanowire Array Micro-LEDs at Telecommunication Wavelengths

InGaAs/InP Quantum Well Nanowire Array Micro-LEDs at Telecom...

引用

2023 Novel Optical Materials and Applications, NOMA 2023

作者： Zhang, Fanlu Su, Zhicheng Li, Zhe Zhu, Yi Gagrani, Nikita Li, Ziyuan Lockery, Mark Li, Li Aharonovich, Igor Lu, Yuerui Tan, Har Hoe Jagadish, Chennupati Fu, Lan Australian Research Council Centre of Excellence for Transformative Meta-Optical Systems Department of Electronic Materials Engineering Research School of Physics The Australian National University CanberraACT2601 Australia School of Electronic Science and Engineering Southeast University Jiangsu Nanjing210096 China Microstructural Analysis Unit University of Technology Sydney SydneyNSW2007 Australia Australian National Fabrication Facility ACT Node Research School of Physics The Australian National University CanberraACT2601 Australia ARC Centre of Excellence for Transformative Meta-Optical Systems Faculty of Science University of Technology Sydney UltimoNSW Australia School of Engineering College of Engineering and Computer Science The Australia National University Canberra CanberraACT2601 Australia

We report the growth, fabrication and characterization of highly uniform p-i-n core-shell InGaAs/InP single quantum well nanowire array light emitting diodes operating at telecommunication wavelengths of ~1.35 and ~1.... 详细信息

关键词： Nanowires

来源：评论

学校读者我要写书评

暂无评论

A Detailed Investigation of Implementation of Internet of Things (IOT) in Cyber Security in Healthcare Sector 2

A Detailed Investigation of Implementation of Internet of Th...

引用

2nd International Conference on Advance Computing and Innovative Technologies in Engineering, ICACITE 2022

作者： Ganai, Prince Tehseen Bag, Akash Sable, Anita Abdullah, Khairul Hafezad Bhatia, Sandeep Pant, Bhasker Shri Jagdishprasad Jhabarmal Tibrewala University Department of Information Technology Rajasthan India Amity School of Law Amity University Chhattisgarh Raipur India University Maharashtra Pune India Universiti Teknologi MARA Perlis Branch Science Laboratory Unit Department of Academic Affairs Arau Campus Malaysia Electronics and Communication Engineering Raj Kumar Goel Institute of Technology U.P. Ghaziabad India Graphic Era Deemed to Be University Department of Computer Science & Engineering Uttarakhand Dehradun India

ISBN: (纸本)9781665437899

The use of 'electronic healthcare technology' is widespread, and it has huge potential to improve 'clinical outcomes and transform' how therapy is administered. On the other hand, medical device and information security is becoming a significant problem. Security breaches were, obviously, an issue even before the advent of digital clinical information. The interdependency of today's documentation, on the other hand, offers various prospective entry points: the ability to obtain from anywhere (whereas earlier, hard copies would have been kept in healthcare institutions and only available through personal breaches);the capacity for security breaches to go unobserved;and access towards a more comprehensive medical record, offering a more precious asset for possible threats (whereas traditionally, medical files may have been divided between many various providers). Previously, a data leak might have affected hundreds or even thousands of people due to missing paper documents or a stolen computer;but, now that this content is digitised and accessible across several networks, a cyber-attack has the capability to harm millions of patients. To demonstrate the concept, celebrity healthcare information has long been a target for cybercriminals. Prior to the introduction of electronic medical records, meanwhile, these vulnerabilities were restricted to hospital workers who had direct access to the documentation. Celebrity health information may now be viewed remotely, thus raising the risk of a data leak. For mitigating this issue, IoT or Internet of Things is widely used in healthcare industry to increase service security. This research has considered secondary qualitative method to gather wide range of information related to the study. © 2022 IEEE.

关键词： Cybersecurity

来源：评论

学校读者我要写书评

暂无评论

Hydrodynamics and the Eigenstate Thermalization Hypothesis

引用

Physical Review X 2025年第1期15卷 011059-011059页

作者： Luca Capizzi Jiaozi Wang Xiansong Xu Leonardo Mazza Dario Poletti Université Paris-Saclay CNRS LPTMS 91405 Orsay France Department of Mathematics/Computer Science/Physics University of Osnabrück D-49076 Osnabrück Germany Science Mathematics and Technology Cluster Singapore University of Technology and Design 8 Somapah Road 487372 Singapore College of Physics and Electronic Engineering and Center for Computational Sciences Sichuan Normal University Chengdu 610068 China Engineering Product Development Pillar Singapore University of Technology and Design 8 Somapah Road 487372 Singapore Centre for Quantum Technologies National University of Singapore 117543 Singapore MajuLab CNRS-UNS-NUS-NTU International Joint Research Unit UMI 3654 Singapore

The eigenstate thermalization hypothesis (ETH) describes the properties of diagonal and off-diagonal matrix elements of local operators in the eigenenergy basis. In this work, we propose a relation between (i) the singular behavior of the off-diagonal part of ETH at small energy differences and (ii) the smooth profile of the diagonal part of ETH as a function of the energy density. We establish this connection from the decay of the autocorrelation functions of local operators, which is constrained by the presence of local conserved quantities whose evolution is described by hydrodynamics. We corroborate our predictions through numerical simulations of two distinct nonintegrable spin-1 Ising models, exhibiting diffusive and superdiffusive transport behaviors. The simulations are performed using dynamical quantum typicality up to 18 spins, for both infinite- and finite-temperature regimes.

关键词： Ising model

来源：评论

学校读者我要写书评

暂无评论

Audio-Based Deep Learning Frameworks for Detecting COVID-19

Audio-Based Deep Learning Frameworks for Detecting COVID-19

引用

European Signal Processing Conference (EUSIPCO)

作者： Dat Ngo Lam Pham Truong Hoang Şefki Kolozali Delaram Jarchi School of Computer Science and Electronic Engineering University of Essex United Kingdom Center for Digital Safety & Security Austrian Institute of Technology Austria Solution Technology Unit Department FPT Software Company Limited Vietnam

ISBN: (数字)9789082797091

ISBN: (纸本)9781665467995

This paper evaluates a wide range of audio-based deep learning frameworks applied to the breathing, cough, and speech sounds for detecting COVID-19. In general, the audio recording inputs are transformed into low-level spectrogram features, then they are fed into pre-trained deep learning models to extract high-level embedding features. Next, the dimension of these high-level embedding features are reduced before fine-tuning using Light Gradient Boosting Machine (LightGBM) as a back-end classification. Our experiments on the Second DiCOVA Challenge achieved the highest Area Under the Curve (AUC), F1 score, sensitivity score, and specificity score of 89.03%, 64.41%, 63.33%, and 95.13%, respectively. Based on these scores, our method outperforms the state-of-the-art systems, and improves the challenge baseline by 4.33%, 6.00% and 8.33% in terms of AUC, F1 score and sensitivity score, respectively.

关键词： COVID-19 Deep learning Sensitivity Neural networks Europe Feature extraction Boosting

来源：评论

学校读者我要写书评

暂无评论

Detection of real-time deep fakes and face forgery in video conferencing employing generative adversarial networks

引用

Heliyon 2024年第17期10卷 e37163页

作者： Sharma, Sunil Kumar AlEnizi, Abdullah Kumar, Manoj Alfarraj, Osama Alowaidi, Majed Department of Information System College of Computer and Information Sciences Majmaah University Majmaah 11952 Saudi Arabia School of Computer Science University of Wollongong in Dubai Dubai Knowledge Park Dubai United Arab Emirates Computer Science Department Community College King Saud University Riyadh 11437 Saudi Arabia MEU Research Unit Middle East University Amman 11831 Jordan Department of Information Technology College of Computer and Information Sciences Majmaah University Majmaah 11952 Saudi Arabia

As facial modification technology advances rapidly, it poses a challenge to methods used to detect fake faces. The advent of deep learning and AI-based technologies has led to the creation of counterfeit photographs that are more difficult to discern apart from real ones. Existing Deep fake detection systems excel at spotting fake content with low visual quality and are easily recognized by visual artifacts. The study employed a unique active forensic strategy Compact Ensemble-based discriminators architecture using Deep Conditional Generative Adversarial Networks (CED-DCGAN), for identifying real-time deep fakes in video conferencing. DCGAN focuses on video-deep fake detection on features since technologies for creating convincing fakes are improving rapidly. As a first step towards recognizing DCGAN-generated images, split real-time video images into frames containing essential elements and then use that bandwidth to train an ensemble-based discriminator as a classifier. Spectra anomalies are produced by up-sampling processes, standard procedures in GAN systems for making large amounts of fake data films. The Compact Ensemble discriminator (CED) concentrates on the most distinguishing feature between the natural and synthetic images, giving the generators a robust training signal. As empirical results on publicly available datasets show, the suggested algorithms outperform state-of-the-art methods and the proposed CED-DCGAN technique successfully detects high-fidelity deep fakes in video conferencing and generalizes well when comparing with other techniques. Python tool is used for implementing this proposed study and the accuracy obtained for proposed work is 98.23 %. © 2024

关键词： Deep conditional generative adversarial networks Deep fake detection Ensemble discriminators

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：